Method
One number out of 100 for each day, built the same way for history and forecast so the two can be compared. Nothing here is a secret formula, and every threshold comes from the record rather than from opinion.
Snow already on the ground sets the day's footing. Fresh snow lifts it towards what is left. Then wind, warmth and rain multiply the result down.
The multiplying matters. A deep base does not rescue a day when the lifts are on wind hold or it is raining on the snow, so those factors have to be able to pull the whole score down rather than shave a little off it. Adding them up would let a good base outvote a ruined surface, which is not how a day on the mountain works.
Fresh snow is a bonus and never a penalty. An earlier version of this took a weighted average of cover and new snow, which quietly marked down a perfectly good day just because nothing had fallen that morning. It now lifts the score when it snows and does nothing when it does not.
Rather than borrow a rule of thumb, the grades are cut from each mountain's own history: every in-season day with a reading, from mid June to the end of September, divided at the 20th, 45th, 75th and 92nd percentile and rounded to the nearest 10 cm. That fixes roughly how often each grade should occur and lets the depths fall where the data puts them.
| Loading… |
The Victorian resorts come out much shallower than the Snowy Mountains, which is the whole argument for doing this per mountain. Seventy centimetres at Falls Creek is the same grade of season as 110 cm at Perisher, and one set of numbers for both would have been wrong for one of them.
The depth score is then bent through those same edges, so an equal score means an equal grade wherever you are: the four edges land on 20, 40, 60 and 80 out of 100. On a day with good weather and no fresh snow, a day's score and its depth grade agree by construction.
We checked the result against the depth guide Michael Paine has published for the Snowy Mountains since 1997, which is the only long-standing scale for these fields. Our independently derived NSW edges land on 60, 110, 150 and 200 cm against his 50, 100, 150 and 200. Identical at the top two, ten centimetres apart at the bottom two. Two different methods arriving at nearly the same place is a good sign for both.
| Warmth | Full marks up to a daily mean of 1°C, then falling away, and hard by 8°C where the surface is slush. A mean of about 1°C with a sub-zero night is an ordinary good day in Australia, so the penalty starts above that rather than at freezing. |
| Cold | Only bites below a minimum of about −12°C, and bottoms out rather than going to zero. Deep cold is uncomfortable here, rarely ruinous, and it is uncommon at these elevations. |
| Wind | Free up to 20 km/h on the modelled series, falling to almost nothing by 45 km/h where lifts start going on hold. |
| Rain | Free up to 3 mm in a day, then falling to about a third of the base. Rain on snow is the fastest way to wreck a surface. |
The wind thresholds are set against the modelled wind we have, not against real ridgetop wind, which is stronger. That is a limitation rather than a choice: the forecast grid cannot see individual ridges. A windy reading here should be read as a floor.
A week is any run of seven days, rolling across the season, so the answer is a real stretch rather than something snapped to a calendar. Windows are ranked on the average daily score, and each one carries the share of days that historically came in good or better, because a week worth booking should be both deep and dependable. The three shown are kept apart from each other so they are genuinely different options.
Forecast days need a depth, and nobody measures the future. Each resort's base comes from its most recent real reading and is carried forward: about 1.2 cm of depth for every centimetre of forecast snow, and about 0.7 cm off for each degree-day above freezing. Both numbers are fitted from 5,883 days of measured daily change in the Victorian archives, then checked against the Spencers Creek record. Typical error is around 2 cm a day, and it compounds, which is why the later days are faded.
A rain term was tested and dropped. Rain and warmth move together closely enough that the fit could not separate them, and the rain coefficient came out with the wrong sign at one resort. Leaving it in would have looked more thorough and been less honest.
The weights are yours to change. A powder skier should not get the same answer as someone who wants a reliable base for a family week. Move the sliders on the front page and the ranking moves with them, including which week comes out best. The presets are just starting points.
It is a guide to natural snow and the weather over it. On the day, the resort's own report is the thing to check.